Abstract
N6-Substituted adenosine analogues containing cyclic hydrazines or chiral hydroxy (ar)alkyl groups, designed to interact with the S2 and S3 receptor subregions, have been synthesized and their binding to the adenosine A1 and A2A receptors have been investigated. Examples of both types of compounds were found to exhibit highly selective binding (Ki in low nM range) to the rat A1 receptor.
Introduction
The physiological functions of adenosine have been extensively studied in recent years. Adenosine exerts its biological effects via extracellular purinergic receptors, termed A1, A2A, A2B, and A3, which are distributed throughout a wide variety of tissues in mammalian systems.1–3 Although adenosine has been approved for clinical use by the FDA for the treatment of supraventricular tachycardia, its therapeutic application is limited by its rapid metabolic inactivation and its nonselectivity for the receptor subtypes.4 There has been considerable interest in the development of adenosine receptor agonists that mimic the pharmacological properties of adenosine but with greater metabolic stability and with higher receptor specificity.5–7 Adenosine agonists with high A1 or A2A receptor selectivity are of potential interest as antihypertensives, antiarrhythmics, analgesics, antipsychotics, and anticonvulsants.8,9 Several recent reports of highly potent and selective adenosine A1 or A2A receptor agonists have focused attention on strategic modifications at the N6-, C2-, and 5′-modified adenosines.5–7,10
However, there have been very few adenosine analogues where a N-N bond exists at the purine 6-position. 6-Hydrazinopurine riboside11 has receptor affinity in the micromolar range (Ki values for A1 = 29.7 μM, A2 = 7.3 μM) but 2-chloro-N6-[4-(phenylthio)-1-piperidinyl] adenosine12 exhibits strong A1 receptor binding and A1 to A2 receptor selectivity (A1 Ki = 0.9 nM, A2 Ki = 470 nM, A2/A1 ratio = 522). The nitrogen isostere of CPA, N6-(1-pyrrolidinyl)adenosine has been reported by us as a potent and selective A1 agonist (Ki = 8.0 nM for A1 and 2800 nM for A2 and a selectivity ratio A2/A1 of 350).10
The model of the N6-region of the A1 receptor has been derived from the structure of (R)-PIA and is based on the assumption that each single part of the C6 substituent (N6, C1, C2, C3 and phenyl) positively contributes to the affinity.13,14 Each of these parts corresponds to a receptor subregion, termed N6, S1, S2, S3, and aryl. The chirality at the C2 carbon, occupying the S2 subregion, produces a high degree of stereoselectivity for R- vs. S- isomers. N6-[(S)-1-Hydroxy-3-phenyl-2-propyl]adenosine, which has a hydroxyl group on the C3 carbon corresponding to the S3 subregion, retains high affinity and selectivity for the A1 receptor (A1 Ki = 2.7 nM, A2 Ki = 390 nM, and a selectivity ratio A2/A1 of 144).13 This paper reports on the design, synthesis, and adenosine receptor binding studies of new N6-substituted adenosine analogues containing cyclic hydrazines or chiral hydroxy (ar)alkyl groups.
Chemistry
Adenosine was used as the starting material in the synthesis of N6-substituted adenosine analogues (3–9). It was acetylated with acetic anhydride, 4-dimethylaminopyridine, triethylamine in acetonitrile at 60 °C (Scheme 1).15 2′,3′,5′-Tri-O-acetyladenosine was converted to the 6-iodo compound 1 by a thermally-induced radical deamination-halogenation reaction with n-pentyl nitrite and diiodomethane in acetonitrile at 60 °C.16 The 6-iodo compound 1 was treated with cyclic hydrazines or chiral (ar)alkylamines in the presence of triethylamine in N,N-dimethylformamide or chloroform/ethanol at 60 °C to provide the N6-substituted triacetates 2a–g, which were subsequently deprotected with sodium methoxide in anhydrous methanol or anhydrous ammonia in absolute ethanol to afford target compounds 3–9.
Scheme 1.

Guanosine served as the starting material in the synthesis of 2-chloro-N6-substituted adenosine analogues (13–18) (Scheme 2). It was acetylated by the same procedure that was used for adenosine but at ambient temperatures followed by reaction with phosphorus oxychloride and N,N-diethylaniline at 60 °C to give 2-amino-6-chloro compound 10 in 87% yield (Scheme 2).17 The key intermediate, the 2,6-dichloro compound 11, was prepared by the radical deamination-halogenation reaction of 2-amino-6-chloropurine riboside 10 in the presence of n-pentyl nitrite and excess carbon tetrachloride.16 The 6-chloro group of 11 was selectively displaced by the cyclic hydrazines or chiral (ar)alkylamines in the presence of triethylamine in DMF or chloroform/ethanol at 60 °C by taking advantage of the greater nucleophilic lability of the 6-position compared to the 2-position. Subsequent deprotection with sodium methoxide in methanol or anhydrous ammonia in absolute ethanol afforded the target compounds 13–18. 2-Iodo-N6-substituted adenosine analogues (21–25) were prepared via the key intermediate, the 6-chloro-2-iodo compound 19, by the same synthetic methodology used for the synthesis of the 2-chloro compounds (Scheme 3).
Scheme 2.
Scheme 3.
Results and Discussion
The A1 receptor affinity and the A2A/A1 selectivity of some chosen analogues were carried out with rat brain or striatal membranes using radioligand binding assays (Table 1). The high A1 affinity and selectivity of N6-(1-pyrrolidinyl)adenosine (3), the corresponding N6-(1-piperidinyl) adenosine (4) and its 2-chloro analogue 14, and the 2-chloro N6-(1-morpholino)adenosine (16) with Ki values for the A1 receptor of 7.3, 3.6, 4.9, and 8 nM, respectively, and with A2A values being in the μM range or higher are of interest. These results, when compared to the low affinity of 6-hydrazinopurine riboside (Ki for A1 = 29.7 ±6.6 μM, Ki for A2a = 7.34 ± 1.12 μM),11 suggest that the destabilizing effects of the polar hydrazino functionality can be offset by stabilizing interactions of larger N6-substituents whose additional carbons interact with the distal hydrophobic N6-subregion. The morpholino analogue 6 is >2000-fold less potent at A1 receptors than the piperidinyl analogue, 4. Thus, the distal ether functionality destabilizes the binding to the receptor. This destabilization can be overcome by adding a 2-chloro substituent as in 16. The chiral compounds synthesized are expected to have interaction with a number of N6-subregions (S1, S2, S3, and aryl) and can be used as probes to study spatial and stereochemical requirements, especially in the S2 receptor subregion. For example, while the (S) and (R) isomers of N6-(1-hydroxy-4-methyl-2-pentyl)adenosines 8 and 9 showed low A1 binding affinity (139 and 224 nM, respectively) and poor A2A affinity (mM range), dramatic differences are seen in the affinities of the (S) and (R) isomers of N6-(1-hydroxy-3-phenyl-2-propyl)adenosines. For example, in the case of 17 and 18, the compound with (S) chirality of the N6-substituent is a factor of about 200 times more potent than the corresponding (R)-isomer at A1 receptors. There is also much greater selectivity between A1 and A2A binding for the (S) compared to the (R) isomer. Related results were obtained for the 2-iodo compounds 24 and 25. The 2-iodo analogues (21, 23, 24, 25) were each less potent at both A1 and A2A receptors than the corresponding 2-chloro analogues (14, 16, 17, 18). Curiously, for the piperidinyl analogues, a 2-iodo group (21) diminished potency at A1 receptors versus the 2-H analogue (4), while for the morpholino analogues, a 2-iodo group (23) enhanced A1 potency versus the 2-unsubstituted compound (6). At the A2A receptors, introduction of a 2-halo group resulted in enhanced potency compared to the 2-H case. The target compounds are stable with respect to deamination by mammalian adenosine deaminase. They are not expected to be substrates for cellular kinases.
Table 1.
Affinities of Selected Adenosine Analogues in Radioligand Binding Assays at A1 and A2A Receptors a,b
| Compound No. | A1 | A2A |
|---|---|---|
| 3 | 8.0 | 280010 |
| 4 | 7.30 ± 1.25 | 29% at 10−4 M |
| 6 | 15,500 ± 1900 | 36±7 % at 10−4 M |
| 8 | 139 ± 44 | <10% at 10−5 M |
| 9 | 224 ± 48 | 30,300±14,800 |
| 14 | 3.55 ± 0.35 | 936 ± 237 |
| 16 | 4.89 ± 0.27 | 1900 ±520 |
| 17 | 2.41 ± 0.46 | 492 ± 87 |
| 18 | 452 ± 65 | 10,400 ± 3600 |
| 21 | 71.0 ± 23.4 | 8630 ± 2220 |
| 23 | 113 ± 35 | 19,400 ±7000 |
| 24 | 23.0 ± 8.0 | 1360 ± 310 |
| 25 | 989 ± 105 | 49 ± 2% at 10−4 M |
Displacement of specific [3H](R)-PIA binding in rat brain membranes, expressed as Ki ± S.E.M. in nM (n = 3–6), or % of displacement at indicated conc.
Displacement of specific [3H]CGS 21680 binding in rat striatal membranes, expressed as Ki ± S.E.M. in nM (n = 3–6), or % of displacement at indicated conc. Radioligand Binding Assays. For all binding experiments, adenosine deaminase was present (3 IU/mL) during the incubation with radioligand. [3H]CGS 21680 binding to striatal A2A-receptors in rat brain was carried out as described18 using 20 μM 2-chloroadenosine to determine nonspecific binding. The binding of [3H]R-PIA to rat cortical A1-receptors was carried out as previously described.19 For competition studies, IC50 values were determined using the Inplot computer program (Graphpad, San Diego, CA) and converted to apparent Ki values using KD values and the Cheng–Prusoff equation.20 KD values in rat brain for [3H]PIA and [3H]CGS 21680 binding were 1.0 and 15 nM, respectively, at A1 and A2A receptors. Concentrations of [3H]PIA and [3H]CGS 21680 used in competition experiments were 1.0 and 5.0 nM, respectively.
In conclusion, N6-substituted adenosine analogues containing cyclic hydrazines or chiral hydroxy (ar)alkyl groups, designed to interact with the S2 and S3 receptor subregions, have been synthesized. Affinity studies of selected analogues to the adenosine A1 and A2A receptors were carried out with rat brain or striatal membranes using radioligand binding assays. Both types of compounds investigated were found to exhibit highly selective binding to the A1 receptor (A1 Ki = low nM range, A2A Ki = > μM range). For pairs of diastereoisomers, the (S)-isomer was significantly more potent than the (R)-isomer. Interestingly, the (S)-isomers of 8, 17, and 24 resemble more closely the structure of (R)-PIA (more active isomer) than (S)-PIA (less active isomer) in terms of alignment of the -CH3 of PIA compared to the -CH2OH of 8, 17 and 24.
Acknowledgments
This research was supported by the NIH. We thank Dr. Jian Zhang for technical assistance.
References
- 1.Williams M. In: Adenosine and Adenosine Receptors. William M, editor. Humana; Clifton: 1990. pp. 1–17. [Google Scholar]
- 2.Gustaffson LE, Wiklund CU, Wiklund NP, Stelius L. In: Purines in Cellular Signalling. Targets for New Drugs. Jacobson KA, Daly JW, Manganiello V, editors. Springer–Verlag; New York: 1990. pp. 200–205. [Google Scholar]
- 3.Linden J. Trends Pharmacol. 1994;15:298. doi: 10.1016/0165-6147(94)90011-6. [DOI] [PubMed] [Google Scholar]
- 4.Pantely GA, Bristow JD. Circulation. 1990;82:1854. doi: 10.1161/01.cir.82.5.1854. [DOI] [PubMed] [Google Scholar]
- 5.Jacobson KA, van Galen PJM, Williams M. J Med Chem. 1992;35:407. doi: 10.1021/jm00081a001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Trivedi BK, Bridges AJ, Bruns RF. In: Adenosine and Adenosine Receptors. William M, editor. Humana; Clifton: 1990. pp. 57–103. [Google Scholar]
- 7.Nair V, Fasbender AJ, Miller LP, Bruce JL. Bioorg Med Chem Lett. 1991;1:481. [Google Scholar]
- 8.Belardinelli L, Linden J, Berne RM. Prog Cardiovasc Dis. 1989;32:73. doi: 10.1016/0033-0620(89)90015-7. [DOI] [PubMed] [Google Scholar]
- 9.Chin JH. Ann Neurol. 1989;26:695. doi: 10.1002/ana.410260602. [DOI] [PubMed] [Google Scholar]
- 10.Nair V, Fasbender AJ. Tetrahedron. 1993;49:2169. [Google Scholar]
- 11.Kusachi S, Thompson RD, Bugni WJ, Yamada N, Olsson RA. J Med Chem. 1985;28:1636. doi: 10.1021/jm00149a016. [DOI] [PubMed] [Google Scholar]
- 12.Knutsen LJS, Lau J, Sheardown MJ, Thompson C. Bioorg Med Chem Lett. 1993;3:2661. [Google Scholar]
- 13.Daly JW, Padgett W, Thompson RD, Kusachi S, Bugni WJ, Olsson RA. Biochem Pharmacol. 1986;35:2467. doi: 10.1016/0006-2952(86)90042-0. [DOI] [PubMed] [Google Scholar]
- 14.van Galen PJM, Leusen FJJ, Ijizerman AP, Soudijn W. Eur J Pharmacol. 1989;172:19. doi: 10.1016/0922-4106(89)90041-2. [DOI] [PubMed] [Google Scholar]
- 15.Matsuda A, Shinozaki M, Suzuki M, Watanabe K, Miyasaka T. Synthesis. 1986:385. [Google Scholar]
- 16.Nair V, Richardson SG. J Org Chem. 1980;45:3969. [Google Scholar]; Nair V, Chamberlain SD. J Org Chem. 1985;50:5069. [Google Scholar]; Nair V, Young D, DeSilvia R. J Org Chem. 1987;52:1344. [Google Scholar]
- 17.Robins MJ, Uznanski B. Can J Chem. 1981;59:2601. [Google Scholar]
- 18.Jarvis MF, Schutz R, Hutchison AJ, Do UH, Sills MA, Williams M. J Pharmacol Exp Ther. 1989;251:888. [PubMed] [Google Scholar]
- 19.Van Galen PJM, Van Bergen AH, Gallo-Rodriguez C, Melman N, Olah ME, Ijzerman AP, Stiles GL, Jacobson KA. Mol Pharmacol. 1994;45:1101. [PMC free article] [PubMed] [Google Scholar]
- 20.Cheng Y-C, Prusoff WH. Biochem Pharmacol. 1973;22:3099. doi: 10.1016/0006-2952(73)90196-2. [DOI] [PubMed] [Google Scholar]


